On Fri, Mar 14, 2008 at 04:08:36PM -0500, Gabriel Sechan wrote:
> 80 column lines-  sure, just don't get religious about it.  If I turn in
>something that has two or three lines in the 81-85 range, don't reject it due
> to a magic number.  If you do, you need serious help on your anal-ness.

I imagine if we were looking at code all day that the word wrap would get
annoying and waste precious nanoseconds to fix.  I think that is the motivation
for <= 80 col.   I can understand why Linus would want that.  That is
one reason why even 1 char over 80 may be a problem.

> 8 space tabs?  No thanks.  Takes up way too much space-  it ends up forcing
> you to either break lines up to an annoying degree, or use shorter, possibly
> less readable variable names.  2-3 works just fine and is no less readable.
> Although I wouldn't reject a code review on it so long as it was a consistent
> indent.  Just don't expect me to write the same way.

I agree line breaks and small variable names are bad.  But there
is a solution to be found in more functions.
Linux uses this and is widely considered to be
beautiful well-written code.   I'm writing a Scheme interpreter at
seberino.org/pyscheme this way.

It reminds me to spawn a new function rather than push the complexity envelope.

> Bite sized functions-  there's a tradeoff there.  Using lots of small
> functions reduces the nesting, which is good for readability.  It also makes
> it harder to figure out what the code is actually doing, which is bad for
> maintainability.  If you have a lot of 1-2 line functions that are only
> called from one spot, you're probably into the territory where smaller
> functions are making your code worse.  If you have them due to the above 80
> columns/8 space tab rule in order to get things to fit, there's no maybe-

I agree lots of 2 line functions would be lame.  Breaking up your logic
with bad function choices is also lame and ruins readability.
You could see it as a programming challenge to try to break up your
code into common sense function encapsulations.  I've *always* been
able to do that.  That would *increase* readability.

cs

-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Reply via email to